From 0354b8721b3b45ff80befdc1252441aa21e95100 Mon Sep 17 00:00:00 2001
From: Martin von Zweigbergk <martinvonz@google.com>
Date: Sat, 16 Oct 2021 23:57:59 -0700
Subject: [PATCH] index: transparently reindex if the index is corrupt

I'm about to change the index format (to remove predecessor
information), which will break the format. Let's prepare for that by
having `IndexStore` reindex the repo if it fails to read the index..
---
 lib/src/index.rs       | 24 +++++++++++++++++++++---
 lib/src/index_store.rs | 18 ++++++++++++++----
 2 files changed, 35 insertions(+), 7 deletions(-)

diff --git a/lib/src/index.rs b/lib/src/index.rs
index 358ea8c27..b8a8f7c92 100644
--- a/lib/src/index.rs
+++ b/lib/src/index.rs
@@ -29,6 +29,7 @@ use blake2::{Blake2b, Digest};
 use byteorder::{LittleEndian, ReadBytesExt, WriteBytesExt};
 use itertools::Itertools;
 use tempfile::NamedTempFile;
+use thiserror::Error;
 
 use crate::backend::{ChangeId, CommitId};
 use crate::commit::Commit;
@@ -219,6 +220,14 @@ impl CommitLookupEntry<'_> {
     }
 }
 
+#[derive(Error, Debug)]
+pub enum IndexLoadError {
+    #[error("Index file '{0}' is corrupt.")]
+    IndexCorrupt(String),
+    #[error("I/O error while loading index file: {0}")]
+    IoError(#[from] io::Error),
+}
+
 // File format:
 // u32: number of entries
 // u32: number of parent overflow entries
@@ -612,7 +621,14 @@ impl MutableIndex {
         persist_content_addressed_temp_file(temp_file, &index_file_path)?;
 
         let mut cursor = Cursor::new(&buf);
-        ReadonlyIndex::load_from(&mut cursor, dir, index_file_id_hex, hash_length)
+        ReadonlyIndex::load_from(&mut cursor, dir, index_file_id_hex, hash_length).map_err(|err| {
+            match err {
+                IndexLoadError::IndexCorrupt(err) => {
+                    panic!("Just-created index file is corrupt: {}", err)
+                }
+                IndexLoadError::IoError(err) => err,
+            }
+        })
     }
 
     pub fn num_commits(&self) -> u32 {
@@ -1389,7 +1405,7 @@ impl ReadonlyIndex {
         dir: PathBuf,
         name: String,
         hash_length: usize,
-    ) -> io::Result<Arc<ReadonlyIndex>> {
+    ) -> Result<Arc<ReadonlyIndex>, IndexLoadError> {
         let parent_filename_len = file.read_u32::<LittleEndian>()?;
         let num_parent_commits;
         let maybe_parent_file;
@@ -1420,7 +1436,9 @@ impl ReadonlyIndex {
         let predecessor_overflow_size = (num_predecessor_overflow_entries as usize) * 4;
         let expected_size =
             graph_size + lookup_size + parent_overflow_size + predecessor_overflow_size;
-        assert_eq!(data.len(), expected_size);
+        if data.len() != expected_size {
+            return Err(IndexLoadError::IndexCorrupt(name));
+        }
         let overflow_predecessor = data.split_off(graph_size + lookup_size + parent_overflow_size);
         let overflow_parent = data.split_off(graph_size + lookup_size);
         let lookup = data.split_off(graph_size);
diff --git a/lib/src/index_store.rs b/lib/src/index_store.rs
index e4e038f7d..00965696e 100644
--- a/lib/src/index_store.rs
+++ b/lib/src/index_store.rs
@@ -26,7 +26,7 @@ use crate::backend::CommitId;
 use crate::commit::Commit;
 use crate::dag_walk;
 use crate::file_util::persist_content_addressed_temp_file;
-use crate::index::{MutableIndex, ReadonlyIndex};
+use crate::index::{IndexLoadError, MutableIndex, ReadonlyIndex};
 use crate::op_store::OperationId;
 use crate::operation::Operation;
 use crate::store::Store;
@@ -54,8 +54,18 @@ impl IndexStore {
         let op_id_hex = op.id().hex();
         let op_id_file = self.dir.join("operations").join(&op_id_hex);
         if op_id_file.exists() {
-            self.load_index_at_operation(store.hash_length(), op.id())
-                .unwrap()
+            match self.load_index_at_operation(store.hash_length(), op.id()) {
+                Err(IndexLoadError::IndexCorrupt(_)) => {
+                    // If the index was corrupt (maybe it was written in a different format),
+                    // we just reindex.
+                    // TODO: Move this message to a callback or something.
+                    println!("The index was corrupt (maybe the format has changed). Reindexing...");
+                    std::fs::remove_dir_all(self.dir.join("operations")).unwrap();
+                    std::fs::create_dir(self.dir.join("operations")).unwrap();
+                    self.index_at_operation(store, op).unwrap()
+                }
+                result => result.unwrap(),
+            }
         } else {
             self.index_at_operation(store, op).unwrap()
         }
@@ -69,7 +79,7 @@ impl IndexStore {
         &self,
         hash_length: usize,
         op_id: &OperationId,
-    ) -> io::Result<Arc<ReadonlyIndex>> {
+    ) -> Result<Arc<ReadonlyIndex>, IndexLoadError> {
         let op_id_file = self.dir.join("operations").join(op_id.hex());
         let mut buf = vec![];
         File::open(op_id_file)